New AI Study Uses Minimal Data to Assess Battery Health and Charge Levels

Lithium-ion batteries have achieved widespread utilization across the globe, energizing mobile devices,gasoline-powered cars, and a diverse range of applications. These batteries stand as the preferred choice for powering our cherished devices. As the shift towards electric vehicles gains momentum, lithium-ion batteries are set to play an important role.

Given the widespread utilization of these batteries, evaluating battery health is paramount to addressing safety concerns associated with emerging battery materials. This becomes crucial due to the limited research into their long-term durability and resilience. Considering their anticipated role in supporting a rising number of vehicles, ensuring effective health assessment methods becomes even more essential.

But,even if one battery fails, it fails the entire battery pack, which disturbs the battery system and may lead to safety issues like smoke, fire, and explosion. Hence, it becomes important to monitor battery states, including parameters like state of charge (SOC) and remaining energy, as well as their statuses, such as overall health condition. 

To tackle this issue, a team of researchers from Carnegie Mellon and the University of Texas at Austin has developed a battery management system to facilitate diagnostics on battery health so that drivers can make informed decisions. They studied the charge curves and used this for battery health estimation and prediction. These curves give maximum capacity that can be used to calculate SOH available battery capacity that can be used to estimate SOC and other energy-related states. The researchers have emphasized that while battery management systems already exist in most electric vehicles, a few qualities make this new model stand out from the rest.

To carry out this research, the researchers studied a total of 10066 charge curves of LiNiO2-based batteries at a constant C-rate. To emphasize this, Jayan, an associate professor of mechanical engineering, said they had a database of around 11,000 experimentally collected charging curves for a particular battery cathode chemistry. They used them to train a machine learning model to predict complete charging curves using sparse data inputs.

This model analyzes only the initial five percent of a battery’s charging process. Using this approach, they can predict how the battery will charge with an incredibly accurate margin of error of just two percent. Impressively, this level of precision is achieved by utilizing a mere 10% of the initial charge curve as input data.

The researchers have said that collecting and using real data as input for the machine learning models will be an important next step to improve the model. Also, the researchers are willing to incorporate environmental variables into the computation of battery charge and subsequent discharge profiles. They are also willing to take data from electric vehicle batteries that are out on the road and explore them. By using actual data from the real world and advanced neural networks, battery management systems can get better at predicting when to charge and discharge batteries. 

Check out the Paper and Reference Article. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our 29k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.

If you like our work, please follow us on Twitter