Electrode Plate Wrinkling Detection During Battery Winding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting electrode plate wrinkling in batteries are plagued by delayed and inaccurate results, leading to increased production costs and resource wastage due to manual observation and disassembly processes.
Innovation Solution
An electrode plate wrinkling detection method utilizing a convolutional neural network-based AI model that acquires and processes images of cells with wrinkled and non-wrinkled electrode plates, constructing a database for training, and performing real-time detection during the battery winding process to identify defective cells and alert personnel for machine adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual observation and disassembly methods are used to detect electrode plate wrinkling, then detection can be performed, but the detection results are lagged and have low accuracy
Solution Approach 1:
The patent replaces the manual mechanical observation and disassembly system with an automated optical imaging system coupled with AI deep learning algorithms. The convolutional neural network processes images of electrode plates to automatically detect wrinkling defects, eliminating the need for manual intervention and achieving both high accuracy and real-time detection capability.
Solution Approach 2:
The patent creates digital copies (images) of the electrode plates and processes these copies through AI algorithms to detect defects. This allows non-contact, high-precision measurement of wrinkling without physically handling or disassembling the actual electrode plates, thereby maintaining detection accuracy while eliminating time delays.
2Reliability
If manual disassembly of cells is performed to determine wrinkled electrode plates, then detection can be achieved, but it causes loss of manpower and material resources
Solution Approach 1:
The patent substitutes manual disassembly operations with an automated image recognition system. The AI model analyzes visual characteristics of electrode plates intact within cells, eliminating the need to physically disassemble cells. This maintains high detection reliability through consistent algorithmic evaluation while completely avoiding the resource waste associated with manual handling and potential damage during disassembly.
Solution Approach 2:
The patent uses optical copying to create image representations of the electrode plates for analysis. This non-invasive approach allows reliable detection of wrinkling defects without physical contact or disassembly, thereby preserving both the integrity of the cells and the resources invested in their production.
3Productivity
If appearance observation of cells is used to detect wrinkled electrode plates, then detection can be performed, but false negatives and false positives occur
Solution Approach 1:
The patent creates detailed digital image copies of the electrode plate surfaces and processes them through convolutional neural networks. These AI algorithms analyze subtle visual patterns and textures that are imperceptible to the human eye, enabling accurate differentiation between actual wrinkling defects and normal surface variations. This eliminates false positives and false negatives while maintaining high detection efficiency through automated processing.
Solution Approach 2:
The patent transforms the detection approach by changing from subjective human visual assessment to objective quantitative image analysis. The AI system evaluates multiple parameters including surface texture, reflectivity patterns, and geometric deviations simultaneously, providing a comprehensive and accurate assessment that eliminates the ambiguities and errors inherent in manual appearance observation.
Data Source
AI summary
Provided are an electrode plate wrinkling detection method and system, a terminal, and a storage medium. The electrode plate wrinkling detection method includes: acquiring images of cells with wrinkled electrode plates and images of cells with non-wrinkled electrode plates; performing type labeling processing on the acquired images and constructing a database by using the labeled images; based on the database, training via a convolutional neural network to generate an electrode plate wrinkling detection model; testing and calibrating the electrode plate wrinkling detection model by using images outside the database; and performing electrode plate wrinkling detection on cell images obtained in a real-time manner during a battery winding process by using the tested and calibrated electrode plate wrinkling detection model.