Ai Can 'hear' When Lithium Battery Is About To Catch Fire

NIST Researchers Develop Sound-Based Method to Detect Lithium-Ion Battery Fires

Researchers at the National Institute of Standards and Technology NIST, including Wai Cheong Andy Tam and Anthony Putorti, have developed a groundbreaking technique to use sound to detect when lithium-ion batteries are on the verge of catching fire.

Lithium-ion batteries are ubiquitous in modern life, powering devices from smartphones and laptops to e-bikes and electric cars. While these batteries are efficient, storing a large amount of energy in a compact space, they can also be hazardous.

When a lithium-ion battery becomes too hot or is damaged, it can enter a chemical reaction known as thermal runaway. To better understand this phenomenon, researchers from NIST collaborated with Xian University of Science and Technology to record the sounds emitted by a lithium-ion battery before and during thermal runaway.

Lithium-ion battery fires are particularly dangerous because they can escalate very quickly. A battery can emit a jet of flame reaching temperatures of up to 1100C - almost as hot as a blowtorch - and it can do so in about a second. This rapid escalation contrasts with traditional residential fires, which usually start more slowly with smoldering or small flames, giving more time for smoke to trigger alarms and warn people.

Furthermore, lithium-ion battery fires dont produce much smoke in the initial stages. By the time the limited smoke activates a conventional smoke alarm, the fire may already be spreading uncontrollably. Therefore, an early detection method is crucial for preventing such fires from becoming deadly.