Unique research using machine learning to predict battery life and fast charging

PhD Student Yizhou Zhang at the Energy & Environmental department at CEVT, Lindholmen.

This Thursday, March 30, Yizhou Zhang will defend his Licentiate thesis Data-driven battery aging diagnostics and prognostics. The research is a collaboration between CEVT and Chalmers University of Technology. The project is partly funded by the Swedish Energy Agency.

The project Data-driven battery aging estimation, which started in 2020, aimed to estimate the Elective Vehicle (EV) battery aging state and predict the remaining useful life of the traction battery, using various data driven approaches and sources. Yizhou has performed his research using machine learning algorithms and battery experiments based on tracking the current and voltage, during battery charging and discharging.

"This research will considerably improve how CEVT can contribute to energy-efficient mobility solutions. We will know how to utilize the costly batteries in a safe and climate-friendly way for a sustainable future. Furthermore, we will also be able to optimize the design and use of energy storage and power systems" says research project leader Dr. John Bergström, Team Leader Advanced Engineering & Industrial PhD at the Energy & Environmental department at CEVT.

"By using data-driven methods combined with customer usage data analysis within laboratory cycle data settings, we can increase the accuracy of both prediction and estimations of remaining useful life of vehicle batteries. This can later be used for fast-charging strategies. It can also be beneficial in second-life applications, like stationary energy storage " says PhD Student Yizhou Zhang at the Energy & Environmental department at CEVT.

This project is connected to CEVT’s strategic objective of developing world-leading mobility offers.  

“Energy storage and consumption in electric vehicles has quite an impact on the overall driving experience. Modelling the physical mechanism of degradation using machine learning and AI technologies could inform OEMs, as well as battery suppliers, which electrical signals are most correlated with aging and state of health of an EV battery. Optimal battery charging protocols are key areas in all CEVT-developed electric platforms, and this is work that provides us with valuable information”, says Shafiq Urréhman, Main architect AI technologies and Manager, Strategic Areas at CEVT.

More information on how to join the defense of Yizhou Zhang Licentiate thesis on Thursday March 30:

Yizhou Zhang, Electrical engineering | Chalmers