Battery Safety for a Zero-Incident Future

Predictive software models are one option when looking for a potentially damaged battery cell needle in a haystack of production modules.

Software can solve challenges related to electric vehicles charging in cold weather conditions. (Sophie Caron/Qnovo)

A fire on January 17 gutted what was reportedly the world’s largest battery storage plant in Moss Landing near Monterey, California. The blaze and concerns over air quality damaged by the toxic smoke caused the temporary evacuation of around 1,500 nearby residents. The facility, owned and operated by Vistra Energy, used lithium-ion batteries manufactured by LG Energy Solution to provide 3 GWh of stored energy to the grid. In a somewhat similar event, on August 1, 2024, an unplugged Mercedes-Benz electric vehicle (EV), reportedly an EQE sedan, caught fire in the basement parking of a residential tower in Seoul, South Korea, sending almost two dozen residents to the hospital for smoke inhalation and damaging dozens of other vehicles in the parking garage.

Lithium-ion batteries are a fundamental pillar of the energy transition from fossil fuel to electricity. (iStock/Qnovo)

Despite incidents like these, lithium-ion batteries are a fundamental pillar of the energy transition from fossil fuel to electricity. They provide the energy storage to power electric transport from scooters to EVs, trucks and railway. They are instrumental in storing renewable solar and wind energy and releasing it at night. Their safety must be a top priority.

After the costly recalls of the Chevrolet Bolt and the Hyundai Kona EVs in 2021 due to battery fires, battery safety climbed to a top priority among EV automakers. The industry’s goal of “zero events” reflects this commitment. Total fire prevention is now the benchmark. In contrast, battery energy storage systems (BESS) increasingly used on the grid remain focused on fire suppression instead of prevention.

Fortunately, technologies are now available to achieve the goal of zero events. One, from my company, Qnovo, is called SentinelX. This technology uses electronic sensing from battery management electronics on board the vehicle with advanced battery models to predict the presence of battery anomalies that may lead to a fire. These predictive models blend sophisticated physical representations of the underlying electrochemical processes with machine-learning models that mine the abundant data collected from the vehicle’s battery.

Battery manufacturing has become highly advanced, making anomalies that lead to catastrophic failures exceedingly and thankfully rare. However, these anomalies do happen and when they do they can have severe consequences. Software solutions are essentially looking for a “needle in a haystack” when it comes to identifying safety issues. Qnovo has collected large data sets related to battery behaviors and failures in the field, including unique cell characteristics from various manufactures. This rate of anomalies is on the order of 10 ppm, meaning ten cells among one million are statistically likely to pose a problem during their lives. Since the battery pack of a typical sedan EV contains between 500 and 5,000 cells, statistically, one in 200 to 2,000 vehicles could exhibit an anomaly that may result in a fire if not addressed early and correctly. As EV deployment accelerates in the next few years, these data points become worrisome.

One primary failure mode that can lead to battery combustion is the slow but dangerous growth of lithium metal dendrites resulting in an electric short within the battery. These dendrites form during the operation of the battery, hence are exceptionally difficult to identify during the manufacturing process. Software intelligence in the vehicle, one of the features offered by SentinelX, monitors the signature of these dendrites, identifies the gravity and progress of the anomaly, then takes the appropriate measures to mitigate or issue an early error code.

SentinelX integrates seamlessly in the software stack of the battery management system and adheres to automotive compliance standards such as ISO 26262.

We applaud the progress that the automotive industry is making towards zero events. Yet, utilities and grid operators of BESS should heed the example of EV makers and prioritize fire prevention in their quest to keeping their communities safe.

Nadim Maluf is CEO of Qnovo.