Can AI help prevent EV battery fires?
Register for free and continue reading
Join our growing army of changemakers and get unlimited access to our premium content
A machine learning system predicts dangerous overheating in lithium-ion batteries
Spotted: As EVs gain popularity around the world, concerns about battery safety are more pressing than ever. One key issue is thermal runaway, which is triggered by an unpredictable spike in temperature that can lead to catastrophic fires and even explosions in lithium-ion batteries. New research from the University of Arizona offers a novel solution to address this, combining machine learning with thermal sensors.
Basab Ranjan Das Goswami, lead researcher and doctoral student, and Professor Vitaliy Yurkiv, project principal investigator, have developed a system that senses, predicts, and identifies thermal runaway events within EV battery cells. Thermal sensors are wrapped around individual cells (of which up to 1000 are closely packed together to form a full battery), and these feed into a machine-learning algorithm that has been trained on historical data.
The algorithm analyses patterns to predict future overheating events, and provides an early warning for potential failures. Das Goswami spoke to Springwise, explaining that “this approach allows for real-time monitoring and early intervention, reducing the likelihood of catastrophic failures in EVs.”
The approach used by Das Goswami and Professor Yurkiv stands apart from traditional methods because they have innovatively combined AI and multiphysics models, along with lightweight sensors. This is more cost-effective than using bulky thermal imaging, and is designed to be integrated into existing battery management systems to accurately predict temperature spikes in real time. As Das Goswami summarises, “this fusion of disciplines enables us to move beyond reactive safety measures to proactive prevention.”
The team secured $599,808 in funding from the US Department of Defense’s Defense Established Program to Stimulate Competitive Research, and the scientists are already working on further refining the algorithm. They are also exploring partnerships with car manufacturers to bring the technology into commercial use.
Written By: Oscar Williams