Lithium-ion batteries are susceptible to thermal runaway, a condition that can occur due to overheating, manufacturing defects, or excessive charging and discharging. This issue can lead to battery failure or even explosions, deterring many from adopting electric vehicles.

Researchers and engineers are working to address this problem by developing materials that can mitigate thermal runaway. Artificial intelligence (AI) has emerged as a powerful tool for enhancing battery safety. By continuously monitoring battery temperature, AI can predict potential overheating risks and enable early detection, allowing for preventive measures.

A notable example of AI’s application in battery safety is the research conducted at the University of Arizona “Advancing battery safety”. The researchers proposed placing temperature sensors on each battery cell, allowing an AI system to monitor temperature fluctuations and predict which cells are likely to overheat.

Researchers considered the structure of electric batteries, which are comprised of numerous interconnected cells. Electric vehicle batteries typically contain around 1000 cells. Overheating in a single cell can propagate to adjacent cells, potentially causing a cascading failure throughout the entire battery.

To account for the complex interactions within a battery pack, the researchers employed a GNN-LSTM model. This model considers the thermal history of each cell and its spatial location within the pack, as the arrangement of cells can significantly influence heat propagation. By training the model on data collected from few battery cells under different conditions, the researchers demonstrated its ability to accurately predict thermal runaway.

The researchers suggest integrating such AI systems into electric vehicle batteries. Training these models on real-world data collected from vehicles can enhance their ability to accurately predict thermal runaway in real-world scenarios.

As the demand for electric vehicles grows, so too does the need for reliable battery technology. By improving battery safety through AI, we can accelerate the transition to a more sustainable future.