Why AI Holds the key to Limiting the Risk of EV Battery Fires - Dr. Umut Genc, Eatron Technologies' CEO

UK-based Eatron Technologies develops AI-powered software platforms for safer, durable and greener batteries in automotive, mobility and beyond. It's also using AI to help vehicle manufacturers limit the risk of electric vehicle (EV) battery fires. 

"EV batteries are typically designed to operate within a ‘buffer’ to help guard against the effects of degradation. But that can mean as much as 5-10% of a battery’s capacity is rendered unusable," Dr. Umut Genc, CEO at Eatron Technologies, tells Auto Futures.

"Our Intelligent Software Layer, which can be deployed on top of an existing battery management system, has the potential to unlock this hidden resource, leading to better performance, an extended lifetime, and increased safety," he adds.

Its battery management software creates a more accurate picture of a battery’s health that makes it possible to safely extend its operational life, in some cases by as much as 25%.

It can also predict EV battery failures before they happen, by identifying degradation mechanisms.

"In reality, EV battery fires are incredibly rare, but our belief is that even one is one too many."

"Vehicle manufacturers now also face the added challenge of regaining the trust of consumers whose opinions of EVs may have been negatively impacted by recent publicity. As an industry we must work to ensure the number of catastrophic battery failures reaches zero and then stays there. Eatron’s intelligent, connected, and safe automotive-grade battery management software has demonstrated that AI holds the key to achieving this," explains Dr. Genc.

Detecting the causes of battery failure is a major challenge, since it often involves a complex combination of factors. One of the most common is lithium plating. This occurs when metallic lithium deposits form around the anode.

"Traditionally, detecting this condition has involved dismantling the battery pack and examining the electrodes, but this is clearly impractical. And while various less invasive techniques have been developed over the years, all have their own limitations, particularly when it comes to implementing them in on-board computers reliably."

By using a technique called feature extraction, Eatron's team transforms the raw data coming from the battery into a unique format that makes anomalies easier to identify.

"By combining this with our own proprietary AI pipeline that accurately captures battery behaviour, our AI diagnostics can predict cell failures before they occur, and with up to 90% accuracy at the module level and zero false positives."

"Once a failure has been detected – or better yet, predicted – the situation can be handled far more effectively. That might mean altering the way the battery is managed to minimise any further degradation, something that might be possible to achieve without minimum loss of performance for the customer," he adds.

AI-Powered ADAS

Eatron has been working with China's NIO since 2020. Their relationship has included projects that go beyond battery management. 

"We implemented an intelligent software stack that continuously changes the suspension damping characteristics according to road conditions, helping to optimise the ride quality. This has been a highly positive and productive partnership for both parties, and together we have successfully demonstrated that Eatron’s intelligent software has an important part to play in the seamless operation of mass-production vehicles," says Dr. Genc,

It also works with OEMs to develop a more natural feel for Advanced Driver Assistance Systems (ADAS).

"AI-powered ADAS are already making an important contribution to efforts to improve road safety. Level 1 and Level 2 technologies such as highway co-pilots and emergency braking systems have long been available, and are now becoming mandatory for all new vehicles."

"As a next step we can implement AI co-pilots, as a driver monitoring system, which can provide feedback and even warnings to drivers in real-time helping them to improve their driving in everyday life. Of course, we need to understand better if this will be acceptable for people in terms of personal privacy and freedom," he says.

Next Stop - North America

Today Eatron has more than 75 employees, and has a R&D centre located in Istanbul. Since its foundation in 2018, it has raised over $13 million in investment, with backing from the likes of VinFast and Germany's Hirschvogel Automotive Group.

"Eatron has established a positive presence in Europe and Asia. North America is the next market for Eatron, and will become our focus over the next two years. In terms of technology deployment, Eatron’s intelligent software will be running on more than one hundred thousand batteries by the end of 2024," says Dr. Genc.

Finally, looking further ahead, we asked him what urban mobility will look like by the end of the decade.

"The adoption of electric vehicles – be it cars, vans, scooters or motorcycles – will continue to accelerate as we move closer to 2030, and this trend is likely to be more noticeable in city areas, where there is a clear benefit in terms of both energy efficiency and air quality. However, it seems that more fundamental changes in mobility are more likely to be realised over a much longer term, and that for the immediate future of 2030, the picture may not be all that different to today," predicts Dr. Genc.