Researchers at the University of Arizona have advanced electric vehicle safety by creating a machine learning model that predicts and prevents lithium-ion battery overheating, addressing a major roadblock in green energy adoption.
Researchers at the University of Arizona have made a significant leap in the quest for safer electric vehicles by developing a machine learning model capable of predicting and preventing dangerous temperature spikes in lithium-ion batteries.
The research, led by doctoral student Basab Goswami, offers a promising solution to one of the most critical safety concerns in the electric vehicle industry. The groundbreaking study, titled “Advancing Battery Safety,” was published in the Journal of Power Sources.
Goswami and his adviser Vitaliy Yurkiv, an aerospace and mechanical engineering professor, crafted a framework that integrates multiphysics and machine learning models to monitor and identify potential thermal runaway in lithium-ion batteries.
“We need to move to green energy,” said Goswami in a news release. “But there are safety concerns associated with lithium-ion batteries.”
Thermal runaway is a hazardous phenomenon where the temperature within a battery rises uncontrollably, potentially leading to fires or explosions. With modern electric vehicles housing over 1,000 closely connected battery cells, the occurrence of thermal runaway in a single cell can trigger a chain reaction, endangering the entire battery pack.
“The temperature in a battery will escalate in an exponential manner and it will cause fire,” added Goswami.
To counter this, the researchers proposed using thermal sensors wrapped around battery cells. These sensors feed historical temperature data into a machine learning algorithm, which then predicts potential thermal runaway events.
“If we know the location of the hotspot (the beginning of thermal runaway), we can have some solutions to stop the battery before it reaches that critical stage,” Goswami added.
Yurkiv expressed his astonishment at the precision of the machine learning model.
“We didn’t expect that machine learning would be so superior to predict thermocouple temperature and location of hotspots so precisely,” Yurkiv said in the news release. “No human would ever be able to do that.”
The new method is an evolution from earlier research that relied on thermal imaging for thermal runaway prediction, which required cumbersome and expensive imaging equipment. This lightweight and cost-effective solution marks a significant improvement.
“Many people are still hesitant to embrace batteries due to various safety concerns,” Goswami added. “To gain widespread acceptance, it’s crucial for the public to know that ongoing research is actively addressing these critical safety issues.”
Goswami’s research represents a significant step toward making electric vehicles safer and more reliable, potentially speeding up the transition to greener energy sources. As this technology evolves and integrates into electric vehicle battery management systems, it could play a monumental role in securing public confidence and fostering broader adoption of electric vehicles.