Automatic quality inspection
of battery cells with AI
The growing electromobility market is placing ever greater demands on the development of high-performance batteries for electric vehicles. With these new requirements, safety and quality requirements are also growing, from the battery cell to the finished battery systems. There are two primary areas of application for battery cell testing in the automotive industry: the testing of battery cells and the testing of battery boxes. The development of new, intelligent solutions for quality testing processes makes it possible to increase the efficiency of battery production and thus reduce costs. AI.SEE™ from elunic offers the opportunity to combine new technology in production with artificial intelligence and establish stable quality processes, optimize throughput and meet the high demands placed on the product.
Quality requirements for battery cells
At the heart of electric vehicles are high-voltage batteries, which are integrated into battery boxes in the vehicle underbody. These high-voltage storage units consist of several battery packs whose battery modules are welded together at the contacts or connected with plug-in modules.
An important process in battery production is welding, which is largely responsible for the quality of high-voltage batteries. Welding processes are mainly used in two areas in the production of cells. Firstly, automated welding processes are used to join the lid and pot of the cell housing. A hermetic seal must be ensured in order to keep the cells airtight and watertight throughout their entire service life. Secondly, contact welds connect the cell body to the poles of the battery. A flawless and precise contact is essential to ensure a powerful current flow in the battery. This makes it all the more important to carry out an automated weld seam inspection during the production of battery cells.
AI.SEE™ offers the possibility to check weld seams for surface defects and continuity.
What tasks can AI.SEE™ solve in eMobility?
- Electrode inspection
- Coating of cells
- Pouch/round cell surface inspection
- Sealing test
- Cell and battery identification
Electrode material is applied in layers to copper and/or aluminum foils. Edge defects and regular shapes must not occur on the complex surfaces. AI.SEE™ can help to reliably detect any embossing defects, especially in the case of organic defects - before cells are wound or stacked, in order to reduce defect costs.
Recognizable example error classes with AI.SEE™:
- Anomalien bei Anoden-/Kathodenbeschichtung
- Testing the separator foil or separator grids
Minor defects in cell coatings can make the cell unsafe or unusable. These organic defects can be made testable using an AI solution by training the neural network. The system learns to automatically evaluate future errors through error labeling (supervised learning).
Recognizable example error classes with AI.SEE™:
- Blowhole
- Bubbles
- Constriction
- Inclusions
- Surface variance
A pouch battery cell takes on a somewhat irregular shape during the degassing process. After the device is formed, the cell pouch is pressed together to iron out and smooth the surface. It is important that pouches are smooth, wrinkle-free and straightened. Cell battery manufacturers use automated inspection systems between these stages to detect surface defects. The complex surface structure of the pouch creates a bumpy, distracting background that can hide wrinkles, bubbles and other defects. The appearance of cell pouches can vary greatly, making it too complicated and time-consuming to search for all defects in detail.
Recognizable example error classes with AI.SEE™:
- Bubbles
- Folds
- Dents
- Other surface anomalies
Stackable or cylindrical cells are filled with an electrolyte after filling. The filling access is sealed with plastic (injection molding) or laser welding. Especially with changing geometries, environmental variables and surface finishes, AI.SEE™ helps to efficiently support chaotic or order-related production without jeopardizing quality requirements.
Recognizable example error classes with AI.SEE™:
- Seam gaps
- Formation error
- Presence
Traceability is an important quality and safety requirement in eMobility, especially in battery production. For traditional machine vision systems, shiny surfaces or changing marking types are often critical decision paths in the overall system design. AI.SEE™ can be trained on different surfaces, for example to make set-up or setup times more efficient in the future or to increase reading reliability.
Recognizable example error classes with AI.SEE™:
- OCR
- OCV
- Codereading
Find out now with just a few questions whether your problem can also be solved by AI.SEE™.
New digital solutions for testing battery cells in Industry 4.0
Reliability:
As the battery production processes described above show, one focus of battery cell testing is the inspection of contact welds and weld seams. Different welding techniques, maximum precision in pole contact and the high number of weld seams pose difficult challenges for conventional quality assurance methods such as visual inspection by employees. The development of new quality control systems makes it possible to address these challenges.
Cost reduction:
AI.SEE™ is a quality control and defect detection system that uses artificial intelligence and scalable camera systems to provide effective solutions for many disciplines of optical battery inspection. Weld inspections and contact inspections in particular are performed with on-demand cameras and lighting systems, optimizing speed and consistency of inspection while reducing the susceptibility to defects due to object or environment variance. In addition, AI.SEE™ can be integrated in-line - i.e. directly into your system - to enable early defect detection before further expensive finishing steps and prevent rework or rejects of higher-value intermediate products or semi-finished products.
Future security:
Compared to human methods of visual inspection, a new, automated quality control in battery cell testing promises high cost savings, more efficient battery production and competitiveness through the use of new technologies.
References
Find out more about the use case.
