Fault Analysis and Effects in Large-Format Lithium-Ion Batteries

The "FAIL" project aims to improve the safety of lithium-ion batteries. For this purpose, a novel CT scanner with a safety chamber is being developed. This will enable the detection of physical processes during thermal runaway, which will enhance our understanding of this phenomenon.

Fraunhofer FFB is manufacturing battery cells with defects, which are then induced into thermal runaway in the diondo device at the Fraunhofer ISE site. The project partner Accure will subsequently analyse this data.

Initial Situation

Global production of lithium-ion cells is rising sharply, whilst even minor manufacturing defects increase the defect rate and significantly heightend the risk of thermal runaway events in vehicles. In the multi-stage cell manufacturing process, analysis and inspection methods are often destructive and too slow to detect critical faults in real time. 

Objective

"FAiL" aims to create a laboratory demonstrator system that visualises critical defects in large-format cells, records the mechanism of their formation in real time, and simultaneously minimises the rejection of flawless cells. The objective is a significant reduction in defects, improved environmental footprint, and a data-driven quality standard for European battery cell manufacturing.

Approach

1. High-speed in-situ CT – A novel liquid-metal X-ray source combined with a photon-counting detector provides 10 µm resolution and a frame rate of ≤ 0.5 ms, making internal short circuits, gas formation, or electrode deformations visible during load cycles – all without destroying the cell.
2. Battery test containment chamber – The measurement environment is housed in a pressure- and temperature-resistant carbon shell that causes virtually no X-ray absorption. This enables the creation of a complete 3D volume reconstruction of a cell up to 50 mm thick – a decisive advancement over conventional thick metal enclosures.
3. Systematic defect framework & AI analysis – Defects (foreign particles, coating irregularities, stacking errors) are specifically introduced into large cells, then subjected to defined electrical loads and simultaneously recorded via in-situ CT and electrochemical measurements. The data forms the training set for AI models that recognize defect patterns, quantify failure risks, and derive precise tolerance limits.
4. Transfer to production – Findings are translated into compact inline CT sensors for production lines. An accompanying monitoring software toolkit evaluates the image and measurement data in real time and immediately triggers alarms in the MES upon detecting critical patterns, enabling defective cells to be sorted out during the production run.

In this way, "FAiL" delivers comprehensive, non-destructive quality and safety monitoring that identifies the cause of thermal runaway and future-proofs European battery cell manufacturing.