DeKIBat

Brief description

The goal of this project is to develop a combined thermographic-optical inspection system for detecting
surface and volumetric defects during the manufacture of electrodes for lithium-ion batteries, with the aim of minimizing scrap in battery cell production and maximizing product quality. To enable robust validation of the system, physical and digitally synthesized defect samples are designed and created. For the first time, the VDA 5.3 standard is being applied to validate the AI-based multimodal imaging system in the production of electrodes for lithium-ion batteries.

The project is divided into a total of seven work packages:

In WP 1, project activities aimed at ensuring the achievement of the described project objectives and the publication of the project results are coordinated. Fraunhofer FFB ensures that the project is carried out efficiently and in a targeted manner.

In WP 2, system requirements are established and critical failure patterns are defined.

In WP 3, physical limit samples are created. Based on the defect catalog, manipulated features (real defects) are generated in production campaigns to serve as test objects for the combined thermographic-optical inspection system.

In WP 4, synthetic limit patterns are designed and created, and defect detection is developed. Based on the definitions of the occurring defects and their characteristics from WP 2, the goal of WP 4 is to detect these defects using modern, AI-based algorithms.

WP 5 involves the development of the thermography system. Based on the hardware and software requirements from WP 2, a prototype demonstrator is built and validated on a laboratory test bench.

WP 6 implements the integration and testing of the individual systems (optical and thermographic) into the production linie, as well as the consolidation of information from the individual systems into a combined inspection method.

In WP 7, the inspection suitability of the systems will be evaluated in accordance with VDA Volume 5.3 by examining both existing and innovative inspection methods, as well as their combinations, for defect detection. Additionally, the uncertainty of the algorithm
will be analyzed, and a comparison of the suitability of different inspection systems will be conducted using synthetic and physical reference test lots.

The findings are beneficial to German OEMs and automotive suppliers, as they enable more cost-effective production while maintaining high quality standards. These points are particularly relevant in the growing battery manufacturing ecosystem in North Rhine-Westphalia. Furthermore, resource savings help reduce dependence on the market.